Deepthi Karkada
A5074115355- Topic Modeling
- Radiomics and Machine Learning in Medical Imaging
- Natural Language Processing Techniques
- Privacy-Preserving Technologies in Data
- Speech and dialogue systems
- Glioma Diagnosis and Treatment
- Brain Tumor Detection and Classification
- Advanced Neural Network Applications
- Multimodal Machine Learning Applications
- Advanced Image and Video Retrieval Techniques
- Speech Recognition and Synthesis
- AI in cancer detection
- Medical Image Segmentation Techniques
- Colorectal Cancer Screening and Detection
- AI in Service Interactions
- Multi-Agent Systems and Negotiation
- Medical Imaging and Analysis
- Machine Learning and Data Classification
Intel (United States)
2018-2023
Although machine learning (ML) has shown promise in numerous domains, there are concerns about generalizability to out-of-sample data. This is currently addressed by centrally sharing ample, and importantly diverse, data from multiple sites. However, such centralization challenging scale (or even not feasible) due various limitations. Federated ML (FL) provides an alternative train accurate generalizable models, only numerical model updates. Here we present findings the largest FL study...
In this work, we quantize a trained Transformer machine language translation model leveraging INT8/VNNI instructions in the latest Intel$^\circledR$ Xeon$^\circledR$ Cascade Lake processors to improve inference performance while maintaining less than 0.5$\%$ drop accuracy. To best of our knowledge, is first attempt industry model. This has high impact as it clearly demonstrates various complexities quantizing We present novel quantization techniques directly TensorFlow opportunistically...
Deep Learning (DL) has the potential to optimize machine learning in both scientific and clinical communities. However, greater expertise is required develop DL algorithms, variability of implementations hinders their reproducibility, translation, deployment. Here we present community-driven Generally Nuanced Framework (GaNDLF), with goal lowering these barriers. GaNDLF makes mechanism development, training, inference more stable, reproducible, interpretable, scalable, without requiring an...
Automatic speech recognition is used extensively in interfaces and spoken dialogue systems. To accelerate the development of new models based on deep learning techniques, developers at Mozilla have open sourced a Speech-To-Text engine known as project DeepSpeech Baidu's research. In order to make model training time quicker CPUs for distributed training, we developed optimizations code scale large number Intel CPU systems, including Horovod framework integration into DeepSpeech. We also...
Abstract BACKGROUND Diffuse astrocytic glioma are common and aggressive malignant primary brain tumors with grim prognosis. Artificial intelligence (AI) has shown promise across predictive, prognostic, diagnostic neuro-oncology applications, towards improving patient management. However, clinical translation deployment hampered by AI models’ requirements for explicit acceleration cards, which not typically considered in environments. Here, we seek the execution of models such...
We introduce a dataset containing human-authored descriptions of target locations in an "end-of-trip taxi ride" scenario. describe our data collection method and novel annotation scheme that supports understanding such locations. Our contains location for both synthetic real-world images as well visual annotations (ground truth labels, dimensions vehicles objects, coordinates the location,distance direction from objects) can be used various language tasks. also perform pilot experiment on...
The present study aims to examine the prevalent notion that people entrain vocabulary of a dialogue system. Although previous research shows will replace their choice words with simple substitutes, studies using more challenging substitutions are sparse. In this paper, we investigate whether adapt speech system when system’s suggested not direct synonyms. 32 participants played geography-themed game remote-controlled agent and were primed by referencing strategies (rather than individual...
Although machine learning (ML) has shown promise in numerous domains, there are concerns about generalizability to out-of-sample data. This is currently addressed by centrally sharing ample, and importantly diverse, data from multiple sites. However, such centralization challenging scale (or even not feasible) due various limitations. Federated ML (FL) provides an alternative train accurate generalizable models, only numerical model updates. Here we present findings the largest FL study...